In producing integrated circuits, it is desirable to provide packaged integrated circuits having plastic or resin packages which encapsulate the die and a portion of the lead frame and leads. These packages have been produced a variety of ways, a few of which will be described here.
Conventional molding techniques take advantage of the physical characteristics of the mold compounds. For integrated circuit package molding applications, these compounds are typically thermoset compounds. These compounds consist of an epoxy novolac resin or similar material combined with a filler, such as alumina, and other materials to make the compound suitable for molding, such as accelerators, curing agents, fillers, and mold release agents.
The transfer molding process as known in the prior art takes advantage of the viscosity characteristics of the molding compound to fill cavity molds containing the die and leadframe assemblies with the mold compound, which then cures around the die and leadframe assemblies to form a hermetic package which is relatively inexpensive and durable, and a good protective package for the integrated circuit.
FIG. 1 depicts the viscosity characteristic curve of a typical mold compound. The Y axis depicts the viscosity of the compound. The X axis represents the time elapsed from a starting point where heat is applied. The mold compound transitions from a high viscosity or hard state to a state where it has very low viscosity after an initial time lapse. The low viscosity stage lasts only a limited time period, typically 20 to 30 seconds, then the compound becomes higher in viscosity and begins to set, or cure. For the entire period the mold compound is heated. The mold compound is thermoset material, so that after being heated for a time period longer than the low viscosity time period it will cure or set.
Transfer molding operations have three stages which correspond to the three phases of viscosity shown in FIG. 1. First there is a preheat stage required to move the mold compound from its hard initial state to the low viscosity state. Second is a transfer stage, where the compound is low in viscosity and easily transported and directed into cavities and runners. This transfer process should be rapid and be completed before the mold compound begins to set. Finally there is a cure stage that occurs following the transfer stage.
FIG. 2 depicts a conventional single plunger transfer mold press 11. The press consists of a plunger or ram 13 that is operated under hydraulic pressure, a top platen 15, a top mold chase 17, a bottom platen 19, and a bottom mold chase 21. A fixed head 23 supports the plunger and a movable head 18 support the top platen, and allows the top platen to be removed for loading and unloading the mold from the top. Mold heaters 25 provide heat to the mold in both the top and bottom platens. An automated mold controller, although not shown, is usually coupled to the press. The top and bottom platens are steel and receive the stresses of the pressing operation, both are heated to provide the temperature needed to perform the transfer molding operation.
FIG. 3 depicts a typical bottom mold chase. In FIG. 3, a top view of bottom mold chase 21 is shown. There are six primary runners 31, each will support a pair of leadframe strips holding wire bonded dies and lead assemblies over each cavity 33. The cavities are formed along the runners 31, which are cylindrical shaped paths that extend from the mold pot 32 and into the rows of cavities. Each cavity is coupled to the runners by a secondary runner 35 which ends in a gate 37, a small opening that lets the mold compound into the cavity. The size and shape of the gate is critical to the speed and control of the transfer and filling stages of the molding process.
FIG. 4 is a detailed drawing of a single runner 31 with a single die cavity 33 shown. The secondary runner 35 is shown coupling the primary runner to the gate 37 and to the die cavity 33. Runner 31 is coupled to the pot 32.
FIG. 5 depicts a cross section BB from FIG. 4. This cross section is taken across the primary runner 31 and along secondary runner 35, and depicts the sloped shape of secondary runner 35 up to the gate 37. The lead frame 51 of a typical bonded part is shown over the bottom mold chase cavity and under the top mold chase cavity 34. Die 53 is shown with the bond wires 55 coupling it to leadframe 51.
The operation of the conventional single pot transfer mold will now be described with reference to FIGS. 2-5. To begin a new molding operation, the mold press is opened and the top and bottom mold chases 17 and 21 are separated. The leadframe and die assemblies are loaded into the bottom mold chases. The mold compound is preheated using an RIF heater or other heater before being placed into the heated mold.
The top and bottom platens are closed, bringing the top and bottom mold chases together. The top and bottom mold chases 17 and 21 are patterned to define a cavity around each die, with the lead frames extending outside the cavity and a space formed around each die. Several leadframe strips each having a row of dies 53 which are bonded to their respective lead frames 51 are placed over the cavities 33 in the bottom mold chase 21. A pellet of resin or similar material mold compound is placed in the mold pot within the top mold chase 17. After an initial heating stage to put the mold compound into its low viscosity state, the plunger or ram 13 is used to begin the transfer phase of the operation. The plunger 13 is brought down through the top mold chase 17 onto the mold compound pellet at a predetermined rate, forcing the mold compound into the primary runners 31. As the runners fill with mold compound the compound will begin filling the secondary runners 35, entering the gates 37 beneath the leadframe and die assemblies 51 and filling the cavities 33.
At the end of the transfer stage the mold compound should fill each cavity 33, preferably at the same time and before the mold compound begins to cure. The rate of the downward force brought by the plunger 13 is varied during the transfer phase to help control the transfer process. Experimental use of the press 11 with a particular mold and compound combination will provide the best combination of pressure and transfer speed which can then be programmed into the automatic press controls to uniformly repeat the process.
After the transfer stage, the packaged parts are cured. Curing the molded parts typically takes 1 to 3 minutes of sitting in the heated mold without disturbance. The compound cure is fairly rapid and may be enhanced by adding curing agents to the compound. At the end of the curing cycle the press is opened and the molded parts and the mold compound sprue or flash in the runners and pot are ejected. This is done by having ejection pins extending through the bottom mold chase 21 and bottom platen 19 push upward under pressure at the same instant, popping the molded parts and sprue out of the bottom mold chase 21. The packaged parts are then removed to other areas where they are separated and trim and form operations performed on the parts.
There are several critical requirements that are to be met in a commercially successful package molding operation. The cavities should be completely and uniformly filled. Using the single plunger mold of FIGS. 2-5 the cavity fill stage is difficult to perform uniformly across such a large mold using the single pot and the long primary runners to transport the mold compound. A problem commonly observed in a single plunger single pot mold operation using a mold such as shown in FIG. 2 is an unacceptable void rate. Voids are areas within the mold cavity that are not filled with compound. These can be areas where the compound fails to flow or where air or other materials are trapped and cause hollow spaces in the packaged part. Voids can be produced if the transfer rate of the mold compound is too slow during the molding process or if air or moisture is trapped in one or more the cavities during the transfer stage.
A second critical requirement is that the wire sweep defect rate be minimized below an acceptable level. Wire sweep occurs as the mold compound enters the cavity through the gates. The mold compound is dense and pulls at the fine wires that couple the bond pads of the die to the leads of the lead frame. These wires will bend under the pressure due to the flow of mold compound. As an example, suppose that in a typical lead frame and die assembly, an average wire sweep of less than 6% is specified. A straight line from the lead frame lead to the bond pad has a sweep of 0%. So if after assembly and mold any wires on a packaged unit are found to have more than 6% sweep, the unit is out of specification, and is considered to be a bad unit. Wire sweep is specified as a maximum allowable parameter and is a big concern in production of integrated circuits, because if the bond wires are moved too much, a wire short between two or more adjacent bond wires often occurs. Alternatively, bond wires sometimes break away. Either condition results in a faulty unit.
Although the wire sweep defect rate which is observed in the single plunger molding presses is adequate for producing low to moderate pin count DIP and flat quad packaged devices, as the device pin counts continue to increase and lead frames become finer in lead to lead pitch, the wire sweep parameter becomes increasingly critical. While it is possible to build 200 pin flat quad devices using these techniques, as the pin count goes towards 400 pins the prior art transfer molding presses using a single mold pot will no longer be economically suitable, due to the low yield and high wire sweep defect rates.
A further disadvantage with a single plunger mold and pellet compound arrangement is that the performance in the two critical areas are inversely dependent on each other. That is, in attempting to perfect the molding process using a single plunger mold, it has been observed that steps taken to reduce wire sweep defects typically increase the void rate, and vice versa. In other words, if the wire sweep defect rate is lowered, the void rate tends to increase. The wire sweep rate can be lowered, for example, by slowing the transfer rate of the mold compound into the cavities. However, doing this tends to increase the void rate. Voids can be reduced by increasing the flow rate into the cavities, but this will tend to increase the wire sweep defect rate.
It has been further observed that the wire sweep and void problems tend to be more severe as the number of cavities and the distance of runners increases. Nonuniform fill can occur along a lengthy runner having many cavities. The cavity closest to the pot will have a faster fill rate than the others. The cavity farthest from the pot will tend to fill at the end of the transfer period, and the rate will be lower because much of the compound has been diverted to other cavities and because the compound is starting to harden. As a result, difficult and time consuming fine tuning of each mold press is required to establish an operation mode which will fill all of the cavities at an acceptable rate, during the low viscosity period, without increasing wire sweep defects to an unacceptable level, particularly for the near and far cavities.
Further, the use of the thermoset molding compound results in a process where the sprue, flash or waste that is left in the pot, the runners and between the devices themselves cannot be reused. Thermoset materials can only be used once in a molding operation, so the excess material must be discarded. Thus the sprue and waste left in the long runners and in the mold pot cannot be recycled.
Also, the conventional molding compound acts as a strong abrasive. During molding, the mold compound is forced out of the mold pot and into the primary runners. The abrasive nature of the mold compound results in rapid wear of the mold pot and the runners, and the plunger or ram itself. This results in expensive rework or replacement of the mold chases on a frequent basis.
An alternative approach for reducing the problems known to the single plunger molding presses of the prior art is to construct a multipellet, multiplunger mold station to replace the single plunger system. A portion of the bottom mold chase of a typical prior art multiplunger mold is shown in FIG. 6. Mold chase 61 has several die cavities 63, grouped in pairs. Runners 67 couple each pair of die cavities to a mold pot 69. Gates 65 enable the mold compound to be transferred into and fill the cavities.
In operation, each of the mold pots 69 receives a so called "mini-pellet" of mold compound. The press is a more complex press than that shown in FIG. 2, and has a plunger for each of the mold pots. The plungers may operate from the top or from underneath the mold. Each mold pot 69 and the short runners 67 act exactly as the single plunger mold of FIG. 2 in operation. The individual plungers are used to start the transfer process, the cavities fill with mold compound as the plunger is pushed into the mold pot, and the transfer phase is completed in a few seconds.
The multiplunger mold process has some advantages over the single pot molding process. The use of the smaller pellets and the shorter runs eliminate the long runners and nonuniform fill times associated with a single plunger press. The pellets used are smaller and therefore do not require preheating, as the mold platens can provide sufficient heat to transition the mini-pellets into the low viscosity state. The wire sweep defect rate can be lowered by providing exact control of the plunger or ram insertion rate, so that the fill is done at a speed which prevents voids while minimizing wire sweep problems. An automated multiplunger press can vary the operation of the plungers during the transfer stage to obtain optimal results.
The nonuniform fill and wire sweep problems associated with the cavities nearest and farthest from the single center pot of the single plunger mold presses are eliminated. Mold compound waste is reduced by the shorter runners.
The disadvantages of the multiplunger molding process are primarily that it requires the use of the mini-pellets. The mini-pellet form of the molding compound is far more expensive per kilogram than the single large pellets used by the single transfer mold. Also, the multiplunger molding station is extremely expensive to manufacture, operate and maintain. The automation of a press with so many plungers is more complex and expensive than the single mold press.
In addition to the added costs, the need for many plungers results in a molding station that has a lower parts per hour throughput than for a conventional single pot mold press. Also the multiple plunger molding system requires complex control and loading and unloading mechanisms. The result is that each station has lower overall throughput than a single plunger mold station, although tighter process control can be achieved. Because the throughput is lowered, additional stations are needed to maintain the same relative level of productivity. High productivity is required to keep the per unit costs low. The need for additional expensive and complex molding stations increases the cost disadvantages for the multiplunger molding systems.
An alternative is a multigang, multipot mold system. This arrangement borrows the simplicity of the single pot mold and adds the multiple pellet idea of the multi-plunger mold by having multiple plungers ganged together and using multiple pots in the mold chase, each feeding two to four cavities with mold compound. By reducing the transfer distance it is hoped wire sweep and void rate problems can be improved. However, tight process control is not available because the plungers all travel at the same speed and pressure unless an external controller is installed. So this alternative has the disadvantages of requiring the mini-pellets of mold compound, while not providing the highly automated process control of a multi-plunger mold system.
Both single plunger and multiplunger mold presses have other disadvantages that are common. The mold compound is an abrasive material. The mold pot and the primary runners receive an abrasive force each time the press is operated. These areas wear quickly and the expensive mold chases must be replaced periodically as a result.
Also, both processes require pelletized mold compound. This material is fairly difficult to produce in the large form, and even more expensive to produce in the minipellet form. The compound is extruded into a rod, which is powdered, and the powder is then pelletized. This is an expensive and complex manufacturing process.
Both pellets and mini-pellets are subject to contamination by moisture and air. It is necessary to perform the molding process under pressure to eliminate trapped air and prevent the formation of voids. Moisture can become trapped in either form of pellet. Moisture contamination of the molding compound can result in additional voids and scrapped devices. Moisture contamination also contributes to package cracking during cure and afterwards to early failure of devices.
U.S. Pat. No. 5,098,626, issued Mar. 24, 1992, and entitled "Method for Packing a Measured Quantity of Thermosetting Resin And Operating a Mold for Encapsulating a Component", and herein incorporated by reference, provides another alternative wherein the mold compound is packaged in individually sealed units. These units each contain mold compound in a quantity needed for a single cavity or pair of cavities for integrated circuit packages. Each of these units is a bag or tube containing mold compound and ending in a bulge or sprout. During molding the bulge or sprout is placed at the end of a runner which feeds a cavity. The sprout is cut and the mold compound is pressed out of the bag into the cavity by individual, multiple plungers.
The '626 patent approach is similar to a multiplunger mold system in that small quantities of mold compound, each of which are individually loaded, are provided. The patent provides a moisture and contamination free packaging system which can be used with an automated loading system. However, like the mini-pellets, many of these bags are required for each run. The abrasion problems are reduced, because the pots and plungers are protected by the packaging. Also, improved uniform fill and reduced wire sweep are possible. But the throughput problems and increased expense for each molding station remain, and the costs for each press are increased further by the added complexity. Also, the packaging of the mold compound in small quantities each in an individual package may lead to an expensive raw material for molding.
Further, the spouted bags of the '636 are fitted into the runner openings. The bags are supplied attached to a tape or spool for continuous feed loading. However, this particular feature of the sprouted bag containers means that the bags and the tape or spool must be custom designed for each particular mold, and if the mold design is changed, a different bag design must be used. Also, the complexity of loading a bag for each cavity or each pair of cavities adds to the precision and cost of the loading equipment used. Further, the molding equipment must include a cutter device for each cavity that cuts the sprout of the bag prior to the transfer stage of molding.
Accordingly, a need thus exists for a transfer molding system which eliminates the problems of the prior art transfer molding systems while retaining a high part throughput rate, low raw material costs, and which is simple to operate, maintain, and uses molding stations that are relatively inexpensive to build. The new system should be compatible with existing single pot transfer mold presses to allow a retrofitting of existing integrated circuit assembly lines. The system should reduce waste of mold compound and reduce the abrasive impact of the mold compound on the equipment used. The new molding system should provide uniform cavity fill and reduced wire sweep defect rates. The system should be general, such that different mold designs can be used with a common mold compound package.